摘要
准确预测主机负载是实现高效动态负载均衡的关键,也是判断负载是否出现异常的重要依据。文章分析了采用时间序列方法进行主机负载预测的不足,提出了基于遗传神经网络的主机负载预测方法,建立了预测模型,并对模型进行了实验评估。
Predicting host load accurately is the key to achieve efficient load balancing,and it is also an important basis for judging whether there are abnormalities. The defects of time series method to predict host load are analyzed. The prediction method based on genetic neural network to predict host load is proposed and the prediction model is established. Also experimental assessment for the model is carried out.
出处
《计算机时代》
2009年第10期12-13,17,共3页
Computer Era
关键词
负载预测
时间序列
遗传神经网络
模型评估
load prediction
time series
genetic neural network
model evaluation
